Published by Associazione Teriologica Italiana Volume 29 (1): 122–127, 2018

Hystrix, the Italian Journal of Mammalogy

Available online at:

http://www.italian-journal-of-mammalogy.it doi:10.4404/hystrix–00052-2018

Research Article

Rensch’s and Bergmann’s Rules in Cis-Andean South-American Howler Monkeys (Mammalia:Alouatta)

Jamile de Moura Bubadué1, George Lucas Sá Polidoro1, Geruza Melo2, Jonas Sponchiado1, Carmela Serio3, Marina Melchionna3,Alessandro Mondanaro3, Silvia Castiglione3, Carlo Meloro4, Pasquale Raia3, Nilton Carlos Cáceres2, Francesco Carotenuto3,∗

1Programa de Pós-Graduação em Biodiversidade Animal, Department of Biology, CCNE, Federal University of Santa Maria, Santa Maria, RS, 97110-970, Brazil2Department of Ecology and Evolution, CCNE, Federal University of Santa Maria, Santa Maria, RS, 97110-970, Brazil

3Department of Earth Science, Environment and Resources, University of Naples Federico II, 80138 Napoli, Italy4Research Centre in Evolutionary Anthropology and Palaeoecology, School of Natural Sciences and Psychology, Liverpool John Moores University, Liverpool, UK

Keywords:Howler monkeysbody sizeRensch’s ruleBergmann’s rule

Article history:Received: e August 2017Accepted: 2 October 2017

AcknowledgementsWe want to acknowledge the editor and two anonymous referees whosecomments helped us in improving the readability of the manuscript.

Abstract

Howler monkeys (genus Alouatta) are large folivorous primates living in South America. We testedfor the application of both Rensch’s rule and Bergmann’s rule to body size variation in Alouatta.We found that Rensch’s rule does apply in howlers. In Alouatta, males exploit dominance rankcompetition, and take advantage from seasonal abundance of high nutritious fruit supply in theirdiet. This mating system and dietary charateristics suggest positive male selection for body size isresponsible for Rensch’s rule. However, since folivory favors large body size in primates (to lowermass specific metabolic rate) and it is the primary dietary habitus in howlers, larger species do occurin the Amazon basin, originating a reversed Bergmann’s rule pattern for both males and femalesat the interspecific level. The spatial and phylogenetic components of such body size patterns ofvariation are both important, implying Alouatta ecomorphological differences to occur above thespecies level, justifying their non-overlapping geographic distribution.

IntroductionIn 1848, Carl Bergmann observed that, among endothermic species,individuals living at high latidudes tend to be larger than those stand-ing closer to the equator (Bergmann, 1847). This was once explainedwith the higher body surface-to-volume ratio in smaller animals, whichhelps dissipating heat in warm habitats (Meiri et al., 2014). The heatconservation hypothesis is not a sufficient explanation for it, and furtherjustifications regard fasting endurance (Lindstedt and Boyce, 1985),environmental predictability (Calder, 1974), and productivity (James,1970).Whatever the reason for Bergmann’s rule is, its application is not as

universal as the term “rule” would suggest (Meiri, 2011). In small-sized animals, like rodents, there are several cases of reverse Bergman-nian pattern (Maestri et al., 2016; Medina et al., 2007; Belk and Hou-ston, 2002; Gohli and Voje, 2016). In the Neotropics, Martinez et al.(2013) recorded a Bergmann’s rule like pattern South to the equatorfor crab-eating fox Cerdocyon, while the reverse applies North to it.These examples suggest that, perhaps unsurprisingly, the relationshipbetween body size and the geography is far more complicated than asimplistic rule would suggest.While Bergmann’s rule describes a latitudinal size cline, Rensch’s

rule predicts that sexual dimorphism (SSD) increases with body sizefor species whose males are larger, and the opposite if females are(Rensch, 1950; Fairbairn, 1997, 2007, 2013; Fairbairn et al., 2007).Male body size is in fact expected to be the primary locus of selectionfor Rensch’s rule, due to male-male competition for mates (Blancken-horn et al., 2006; Gordon, 2004).Since Bergmann’s rule predicts larger body size with latitude, and

Rensch’s rule predicts larger SSD with males larger than females, the

∗Corresponding authorEmail address: f.carotenuto@ymail.com (Francesco Carotenuto)

effect of the two patterns may conflate, provided the largest speciesoccur farther from the equator (Eweleit and Reinhold, 2014; Werneret al., 2016). Thus, under Rensch’s rule, the latitudinal trend in malebody size may steepen (Blanckenhorn et al., 2006).

In primates, both Bergmann’s and Rensch’s rules were explored anumber of times (Gordon, 2004; Clauss et al., 2013). Harcourt andSchreier (2009) found support for Bergmann’s rule, and Smith andCheverud (2002) found Primate as a whole to obey Rensch’s rule. Yet,when the model is controlled for the phylogeny, the relationship disap-pears for both Platyrrhini and Strepsirhini.

Howler monkeys (genus Alouatta) are an ideal study model to testBergmann’s rule, Rensch’s rules, and their potential interaction. Howl-ers are highly sexually dimorphic (Ford, 1994), and widely distributedin South America. Alouatta belongs to the Atelidae family. The genuscomprises 11 species, which diversified during theMiocene, when theircommon ancestor expanded its geographical range through the An-dean Cordillera (Meloro et al., 2014a; Lynch-Alfaro, 2012). Biogeo-graphically, there are two distinct, monophiletic groups of howlers.Trans-Andean Alouatta include species distributed over Central Amer-ica and Trans-Andean Colombia and Ecuador. Cis-Andean Alouattainclude the South American species (Cortés-Ortiz et al., 2003). Al-though widely distributed, most Alouatta species are restricted to asingle biome and show little geographic overlap with each other (i.e.they tend to be parapatric). Howler monkeys are highly-specializedleafs feeders. As withmany folivorous taxa, thesemonkeys tend to havea low activity pattern as compared to other South-American primatessuch as capuchins (Cortés-Ortiz et al., 2003; Lynch-Alfaro, 2012).

We tested whether Rensch’s and Bergmann’s rules apply to Alouattaspecies and their interaction. We focused upon the Cis-Andean cladewe have studied in the field. This is welcome because only Cis-AndeanAlouatta occurs outside the Tropics, and occupy, as a group, a muchwider latitudinal range than the Trans-Andean clade, making them bet-

Hystrix, the Italian Journal of Mammalogy ISSN 1825-5272 20th June 2018©cbe2018 Associazione Teriologica Italianadoi:10.4404/hystrix–00052-2018

Rensch and Bergmann’s rules in howler monkeys

ter suited to study latitudinal effects on body size variation. We usedlatitude as the predictor variable in both cases, but since latitude is justa proxy for environmental variables (see Martinez et al., 2013; Maestriet al., 2016), such as temperature, precipitation and vegetation type, wefurther tested for the impact of these variables. Specifically, we statedthree explicit hypotheses:

1. Alouatta species follow Rensch’s rule. In these primates, malestend to be larger than females and compete with each other(Meloro et al., 2014b). Thus, we expect a stronger relationshipbetween sexual dimorphism and the size of males rather than thesize of females (i.e. male-driven increased SSD with size).

2. Alouatta species follow Bergmann’s rule at the interspecific level(Pincheira-Donoso, 2010; Meiri, 2011).

3. Sexual size dimorphism (SSD) varies with the latitude. Thishypothesis follows form hypotheses 1 and 2. If Rensch’s andBergmann’s rule both apply in Alouatta, then sexual dimorphismwill also correlate with latitude (as well as with the environmentalvariables latitude is a proxy for).

Materials and Methods

We collected data for 227 skulls of Alouatta (Tab. 1), belonging tothe following six different species, A. belzebul, A. caraya, A. guar-iba, A. macconelli, A. nigerrima and A. seniculus (with the exclusionof A. sara because of the lack of specimens in the museums we vis-ited) housed in themain Brazilianmuseums: MuseuNacional (MNRJ),Museu Paraense Emílio Goeldi (MPEG), Museu de Zoologia da Uni-versidade de São Paulo (MZUSP), Museu de História Natural Capãoda Imbuia (MHNCI), Coleção Científica de Mastozoologia da UFPR(DZUP), Museu de ciências naturais da Fundação Zoobotânica do RioGrande do Sul (MCN/FZB). We included only specimens for whichcollection locality geographical coordinates were available. Unfortu-nately, body size data were not reported in most cases. We there-fore relied on geometric morphometrics techniques to retrieve size in-formation from the collected specimens. In geometric morphometrics,landmarks (placed on specimens’ anatomically homologus points) areplaced all along the structure of interest (in this case the skull).

The specimen centroid size (a measure of the size of the landmarksconfiguration) is a very good proxy for body size (Zelditch et al., 2012).The data-acquisition protocol includes taking skull photographs takenat a fixed distance (1 m) to the digital camera applying zoom to cor-rect possible deformations due to lenses (Meloro et al., 2008). Then,digital photographs were landmarked by a single investigator (N.C.), inorder to prevent inter-observer error, using the software tpsDig2 2.16(Rohlf, 2015). When taking photos, we positioned a scale bar adjacentto the specimen in order to transform digital pixels into linear measure-ments, allowing us to compute skull size directly from the configura-tion of landmarks. Twenty-three homologous landmarks were identi-fied and digitized in order to extract skull size information, in the formof the natural logarithm of centroid size (LnCS, see configuration oflandmarks used at Meloro et al., 2014b). The protocol concluds withanalytical and geometric transformation that reduce acquisition errorand scales all the measured specimens to the unity (Rohlf and Slice,1990).

Figure 1 – Map of South America showing the geographic distribution of Alouatta speci-mens. Sampling localities of di�erent species and sexes are shown by di�erent symbols.

In order to study the geographical patterns of species body sizeand SSD, we collated geographically the specimens by performing aclassic spatial sampling protocol. We overlaid the geographic dataset(sampling points) with a grid and then computed mean female bodysize, mean male body size, and SSD per species per each cell of thegrid. This way, each body size mean and SSD datapoints acquire thegeographical coordinates of the cell centroid they belong, separatelyfor each species. We used a grid with a 250×250 km cell resolutionin order to maximize the number of useful cells as to have at least oneindividual of both sexes for each species in a cell. In the end, the ori-ginal dataset reduced from 227 specimens to 82 samples distributed in38 total useful cells following the criteria explained above (Fig. 1; seealso Fig. S1 and Tab. S2).

Environmental variables

For each specimen, we recorded the geographic coordinates of its col-lection locality and a set of four related environmental variables: An-nual Mean Temperature (BIO1), Temperature Seasonality (BIO4), An-nual Precipitation (BIO12) and Precipitation Seasonality (BIO15) (Hij-mans et al., 2005). These variables are provided as geogrphical ras-ter grids at 50×50 km cell resolution (WorldClim raster database,worldclim.org). Temperature and precipitation, together with their vari-ability, determine the dominant climate of a region. Two additional

Table 1 – Skull sample size for the six Alouatta species considered in this study. The data reported are referred to the whole sample of specimens and the reduced dataset after thespatial sampling by cell grid.

#Specimens #Females #Males Average males Average femalesSpecies #Specimens #Females #Males in cells in cells in cells per cell per cell

Alouatta belzebul (Linnaeus, 1766) 65 36 29 16 8 8 1 1Alouatta caraya (Humboldt, 1812) 44 19 25 22 11 11 1.1 1.1Alouatta guariba (Humboldt, 1812) 47 19 28 18 9 9 1 1Alouatta macconelli Elliot, 1910 11 5 6 6 3 3 1 1Alouatta nigerrima Lönnberg, 1941 10 5 5 2 1 1 1 1Alouatta seniculus (Linnaeus, 1766) 50 29 21 20 9 9 1 1

Total 227 113 114 84 42 42

123

Hystrix, It. J. Mamm. (2018) 29(1): 122–127

Figure 2 – Box plot with standardized deviation of natural log transformed centroid size(LnCS) across sexes. Black string: median, white box: first interquartile, bar: secondinterquartile. Di�erent species and sexes are shown by di�erent symbols.

variables were taken from the Atlas of the Biosphere [net primary pro-ductivity (NPP) and evapotranspiration, https://nelson.wisc.edu/sage/data-and-models/atlas/maps.php], by using DIVA-GIS 7.5 software(http://www.divagis.org/download). These variables are informative asper the energy (biomass) available to species, which may impact onsexual dimorphism in primates (Plavcan, 2012).

Statistical analyses

First, we used the cell-averaged LnCS to test for differences in sex andspecies (and their interaction) by using a two-way ANOVA. To test theexistence of Rensch’s rule, we computed the Sexual Size Dimorphism(SSD) for each species in each cell as the difference between male andfemale LnCS and used it as response variable versus female and maleLnCS as covariates in a linear regression. Then, to test for Bergmann’srule, we used the cell-averaged female and male mean LnCS for eachspecies against latitude of the cells centroids. Similarly, to test for apotential role of climate on these species skull size variability, we ranregression models including environmental variables as predictors andthe sex-averaged LnCS for each species in each cell as response. Theinteraction between Rensch’s and Bergmann’s rule was tested by us-ing SDD per cell as the response variable, and the latitude of the cellcentroid, plus environmental variables in separate regression models(one for each predictor).

Controlling for the spatial autocorrelation and phylogenetic re-latedness

When dealing with geographically distributed variables, their spatialautocorrelation must be accounted for (Diniz-Filho et al., 2003). Tothis aim, we computed Moran’s Index on both cell averaged SSD andmale and female LnCS by using the software Sam v.4.0 (Rangel et al.,2010). We anticipate here we found significant spatial autocorrelationin the model’s residuals (Table S4), hence we took it into account inour analyses by including a new set of variables describing the spatialstructure of the predictors. This is done by performing the Eigenvector-based Spatial Filtering (Griffith, 2013), which is a method that uses adistance or connectivity matrix to perform a Principal Coordinate Ana-lysis (PCOA). Then, the method selects the eigenvectors iteratively asto minimize spatial autocorrelation in the residuals (Griffith and Peres-Neto, 2006). The algorithm starts by using the eigenvectors as explan-atory variables in an Ordinary Least Square (OLS) regression with thetrait (here cell averaged male, female LnCS, or SSD, alternatively) asthe response variable. The residual autocorrelation is computed and theeigenvector in the model with smallest Moran’s I coefficient is selec-ted and becomes fixed. The algorithm proceeds iteratively by adding

new eigenvectors in the (multiple) regression until the residuals auto-correlation is below a given threshold for p-values, usually 0.05 (Diniz-Filho et al., 2012; Carotenuto et al., 2015). Once the algorithm findsthe most relevant eigenvectors, we can include them as additional co-variates (herein named “spatial filters”) in the regression models. Thealgorithm described above was performed by using the software SAM(Rangel et al., 2010).

Due to species shared ancestry, we also needed to take into accountpossible phylogenetic effects. We used as a reference the Alouatta treeprovided by Cortés-Ortiz et al. (2003). The tree was trimmed to ourdataset (i.e. by including Cis-Andean clade species only) using theMesquite 2.75 software (Maddison and Maddison, 2011) (Fig. S3).We excluded A. nigerrima from the phylogenetic analyses because ofits unstable phylogenetic positioning. Branch lengths were based onthe estimated minimum ages, as reported in Cortés-Ortiz et al. (2003).The ages of undated nodes were estimated using the BLADJ algorithm(branch length adjustment; Webb et al., 2008) in the Phylocom soft-ware. Since specimens were used as our sample base for the phylo-geny, polytomies within each species were employed when more thanone specimen per species was in the tree, conventionally setting tipswithin species at 0.1 Ma. The inclusion of multiple specimens per spe-cies is particularly important here, since potential within-species vari-ation related to sex, geographical distribution and climate are the fo-cus of the present paper. The multispecimens phylogenetic regressionswere performed applying phylogenetic generalized least squares regres-sions (Ives et al., 2007), between environmental variables and the cellaveraged values of SSD, of male LnCS, and of female LnCS, respect-ively, while accounting for interspecific variability, using the functionpgls.SEy in “phytools” (Revell, 2012).

We performed all the regressions in four ways: by using OrdinaryLeast Squares (OLS); OLS with the spatial filters as additional covari-ates to account for spatial autocorrelation; performing PGLSs to ac-count for phylogenetic relatedness; and drawing a more complex setof models by performing PGLS regressions including spatial filters asadditional covariates to account for both phylogenetic relatedness andspatial autocorrelation at the same time.

ResultsBy grouping specimens using the 250×250 km cell resolution grid weidentified 38 cells. Where a species was present with individuals ofone sex only it was excluded. By this criterion, the number of cellsavailable to testing reduced to 34.

In the two-way ANOVA model using species and sex as factors,we found size to be significantly different for both factors (Species:F=15.626, df=5, p<0.001; Sex: F=392.251, df=1, p<0.001), with nointeraction between them (F=0.801; df=5, p=0.553). Males are largerthan females in all species, with A. macconnelli and A. seniculus beingthe largest overall (Fig. 2).

Hypothesis 1. Rensch’s RuleWe found strong evidence in favour of Rensch’s rule (Tab. 2, Fig. 3).Males skull size is significantly related to SSD, the same applies underPGLS, and when spatial filtering is applied. No significant result wasfound for females (Tab. 2, Fig. 3).

Hypothesis 2. Begmann’s ruleAgainst hypothesis 2, we found the reverse of Bergmann’s rule to applyto both females and males in Alouatta when using the Ordinary LeastSquares regression model (Tab. 3, Fig. 4). The slope is positive, whichmeans a decrease of males and females’ LnCS southward. For males,the same applies when accounting for spatial and phylogenetic effects(Tab. 3). For females, Bergmann’s rule disappeared under PGLS, andunder PGLS plus spatial filter (Tab. 3). As regards the relationshipbetween males LnCS and the environmental variables we found thatwhen considering the BIO1 as covariate, all the four models were pos-itive but significant only with the OLS and the OLS plus spatial filter(Tab. S5). When we considered the BIO4 as predictor, the model was

124

Rensch and Bergmann’s rules in howler monkeys

Table 2 – Results of regressions between body size and the degree of sexual dimorphismin Alouatta, performed separately for males and females, respectively. male LnCS = naturallogarithm of males’ centroid size, female LnCS = natural logarithm of females’ centroidsize, SSD = sexual size dimorphism. The specification “PGLS” indicates phylogeneticgeneralized least squares regression results. The specification “s.filter” indicates spatialfiltering was imposed on the regressor to account for spatial autocorrelation.

Slope s.e. t p logLik

male LnCS vs SSD 0.843 0.181 4.659 <0.0001 62.701male LnCS vs SSD+ s.filter 0.830 0.177 4.689 <0.0001 64.135

male LnCS vs SSDin PGLS 0.588 0.112 5.242 <0.0001 68.288

male LnCS vs SSD+ s.filter in PGLS 0.601 0.114 5.286 <0.0001 66.316

female LnCS vs SSD −0.156 0.185 −0.845 0.403 61.873female LnCS vs SSD+ s.filter −0.169 0.181 −0.937 0.355 63.291

female LnCS vs SSDin PGLS −0.323 0.171 −1.893 0.066 59.911

female LnCS vs SSD+ s.filter in PGLS −0.331 0.176 −1.882 0.068 57.861

male LnCS vs female LnCS 0.807 0.151 5.355 <0.0001 64.890male LnCS vs female LnCS+ s.filter 0.784 0.157 4.983 <0.0001 65.068

male LnCS vs female LnCSin PGLS 0.382 0.121 3.164 0.003 65.573

male LnCS vs female LnCS+ s.filter in PGLS 0.368 0.122 3.010 0.005 63.672

always positive and significant for all the models. All the models werenegative and significant when considering BIO12 as predictor, whereasno model was significant when considering BIO15. The relationshipsbetween male LnCS and evapotranspiration were all positive and sig-nificant, whereas no significant result was found when considering netprimary productivity (Tab. S5).

For females, the relationship between LnCS and BIO1 was signific-ant and positive only when considering the spatial information. Therelationships beteween females LnCS and BIO4 were negative and sig-nificant only for the OLS and the OLS + spatial filter models, and thesame applied when considering BIO4 except for the sign of the slope.The relationship between BIO12 and females LnCS was positive andsignificant only for the OLS and OLS + spatial filter, whereas no sig-nificant relationships were found for BIO15. Evapotranspiration waspositive and significant for the first two models (Tab. 3), whereas nomodel was singnificant when considering net primary productivity aspredictor (see Tab. S5).

Figure 3 – Regression plots for Rensch’s sexual size dimorphism and female and malenatural log transformed centroid size (LnCS). Species and sexes are labelled by di�erentsymbols.

Table 3 – Results of regressions between body size latitude in Alouatta, performed separ-ately for males and females, respectively. male LnCS = natural logarithm of males’ centroidsize, female LnCS = natural logarithm of females’ centroid size, Latitude = latitude of thegrid cell in decimal degrees. The specification “PGLS” indicates phylogenetic generalizedleast squares regression results. The specification “s.filter” indicates spatial filtering wasimposed on the regressor to account for spatial autocorrelation.

Slope s.e. t p logLik

Latitude vs male LnCS 84.723 21.199 3.997 <0.0001 −135.817Latitude vs male LnCS+ s.filter 85.950 22.112 3.887 <0.0001 −135.788

Latitude vs male LnCSin PGLS 0.002 0.001 2.298 0.027 51.671

Latitude vs male LnCS+ s.filter in PGLS 0.002 0.001 2.104 0.042 49.722

Latitude vs female LnCS 88.925 27.315 3.256 0.002 −137.903Latitude vs female LnCS+ s.filter 90.029 28.637 3.144 0.003 −137.891

Latitude vs female LnCSin PGLS 0.001 0.001 1.423 0.163 57.165

Latitude vs female LnCS+ s.filter in PGLS 0.001 0.001 1.119 0.271 55.322

Table 4 – The degree of sexual dimorphism (SSD) regressed against latitude in Alouatta.Regressions were performed separately for males and females, respectively. lat = latitudein decimal degrees. The specification “PGLS” indicates phylogenetic generalized leastsquares regression results. The specification “SF” indicates spatial filtering was imposedon the regressor to account for spatial autocorrelation.

Slope s.e. t p logLik

Latitude vs SSD 47.288 34.289 1.379 0.176 −141.838Latitude vs SSD+ s.filter 46.432 34.649 1.340 0.189 −141.672

Latitude vs SSDin PGLS 0.001 0.001 1.125 0.268 53.671

Latitude vs SSD+ s.filter in PGLS 0.001 0.001 1.202 0.237 51.760

Hypothesis 3. Sexual size dimorphism and latitudeThere is no significant relationship between the degree of sexual sizedimorphism and latitude, irrespective of whether spatial autocorrela-tion, or phylogeney are accounted for (Tab. 4). The same applies withenvironmental variables (see Tab. S5).

DiscussionThe body size of individuals within species can be shaped by envir-onmental (Bergmann’s rule), ethological, or ecological factors, likecharacter displacement, or the mating system (Bubadué et al., 2016;Carotenuto et al., 2015; De Lisle and Rowe, 2015; Meiri et al., 2014;

Figure 4 – Regression plots for Bergmann’s rule on its original form, latitude, and femaleand male natural log transformed centroid size (LnCS). Species and sexes are labelled bydi�erent symbols.

125

Hystrix, It. J. Mamm. (2018) 29(1): 122–127

Shuster and Wade, 2003; Lande, 1980). The way individuals of bothsexes within a species react to these drivers over the evolutionary timedetermines the degree of sexual dimorphism, and how it unfolds overspace.South American howler monkeys are folivorous primates. They are

large, which helps food digestion and lowers mass specific metabolicrates (Meloro et al., 2014a; Cáceres et al., 2014) as compared to otherSouth-American primates, such as capuchins (Cáceres et al., 2014;Canale et al., 2009; Fragaszy et al., 2004). Howler monkeys show dom-inance rank competition between males (Kay et al., 1988) meaning theintensity of male/male context over mates is strong, which promotessexual dimorphism (Kelaita et al., 2011; Plavcan and Van Schaik, 1997;Ford, 1994). In general terms, folivory and arboreality correlate to littlesexual size dimorphism in primates (Plavcan and Van Schaik, 1997),but Alouatta possibly makes an exception (Plavcan and Van Schaik,1997; Ford, 1994). Competition takes place between Alouatta spe-cies (Peres, 1997), meaning the scope for sexual dimorphism is poten-tially counterbalanced by interspecific competition pressure (so far assize overlap between species is minimized to avoid competition, Dayanand Simberloff, 2005). However, dietary differences between sexes arenegligible in Alouatta species (Pavelka and Knopff, 2004; Glander andTeaford, 1995; Bicca-Marques and Calegaro-Marques, 1994) meaningthere is little competition for food between males and females. There-fore, the positive relationship between male size and sexual size di-morphism we found (in keeping with Rensch‘s rule) must be driven bymale/male interactions, at least to some extent. Ravosa and Ross (1994)found evidence for Rensh’s rule in Alouatta, and similarly related theirfindings to the prolonged growth of males in this genus. It has beensuggested that an even distribution of resources through the year de-creases sexual dimorphism in polyginous species (Isaac and Johnson,2003). As Alouatta experience a seasonal abundance of fruit in theirdiet (Bicca-Marques and Calegaro-Marques, 1994; Peres, 1997), it ispossible that males are better in securing this occasional resource sur-plus than females, which would burst their growth (Weckerly, 1998)and help intrasexual competition over mates. We found that Brownhowler monkeys A. guariba follows Bergmann’s rule. It is interestingnoticing that the percentage of leaves in the diet of the brown howlerdecreases with latitude in Belize (Chaves and Bicca-Marques, 2013).Assuming this to be true for other species as well, it suggests that fo-livory decreases body size differences with latitude within species, butincreases it between species. This would help explaining why we foundevidence for a reverse Bergmann’s pattern for both males and females(Tab. 4), and why larger species do occur in the Amazon basin (Fig. 1).In summary, our results indicate that body size variation in Alou-

atta follows Rensch’s rule. A possible explanation of such a patterncan be addressed to the Howler monkeys’ dominance rank competitionmate system (Kay et al., 1988) that, coupled with the seasonal abund-ance of fruits supply in the Amazon basin, favours selection for largesized males in equatorial species. We found a reverse Bergmann’s rulepattern between species, although Bergmman’s rule may be still validwithin some individual species. This possibly depends on the relativeconsumption of leaves versus fruit in the diet, which is higher in theAmazon basin. Whereas larger howlers are folivorous, the occasionalinclusion of fruit in the diet may increase body size within species, es-pecially in males.

Limitation of the studyWe urge the reader to consider that the results we found are valid forsome one half of the living Howler species. While this does not weakenthe validity and the soundness of our findings, it would be interestingto explore, in the future, whether the same patterns accrue to Trans-Andean howlers.

ReferencesAndersson M., 1994. Sexual Selection. Princeton University Press, Princeton, NJ.Ashton K.G., Tracy M.C., de Queiroz A., 2000. Is Bergmann’s rule valid for mammals?

American Naturalist 156: 390–415.Badyaev A.V., Hill G.E., Whittingham L.A., 2002. Population consequences of maternal

effects: sex-bias in egg-laying order facilitates divergence in sexual dimorphism betweenbird populations. Journal of Evolutionary Biology 15: 997–1003.

Belk M.C., Houston D., 2002. Bergmann’s Rule in Ectotherms: A Test Using FreshwaterFishes. The American Naturalist 160(6).

Bergmann C. 1848. Über die Verhältnisse der Wärmeökonomie der Thiere zu ihrer Grösse.Göttingen. [in German]

Bicca-Marques J.C., Calegaro-Marques C., 1994. Activity budget and diet of Alouattacaraya: an age-sex analysis. Folia Primatol 63: 216–220.

Bidau C.J., Martinez P.A., 2016. Sexual size dimorphism and Rensch’s rule in Canidae.Biological Journal of the Linnean Society 1–15.

Blackburn T.M., Gaston K.J., Loder N., 1999. Geographic gradients in body size: a clari-fication of Bergmann’s rule. Diversity and Distributions 5: 165–174.

Blanckenhorn W.U., Stillwell R.C., Young K.A., Fox C.W., Ashton K.G., 2006. WhenRensch meets Bergmann: does sexual size dimorphism change systematically with lat-itude? Evolution 60(10): 2004–2011.

Blomberg S.P., Garland T., Ives A.R., 2003. Testing for Phylogenetic Signal in ComparativeData: Behavioral Traits are More Labile. Evolution 57: 717–745.

Bonvicino C.R., Lemos B., Seuánez H.N., 2001. Molecular phylogenetics of howler mon-keys (Alouatta, Platyrrhini). Chromosoma 110: 241–246.

BrownW.L., Wilson E.O., 1956. Character displacement. Systematic zoology 5(2): 49–64.Bubadué J., Cáceres N., dos Santos Carvalho R., 2016. Ecogeographical Variation in Skull

Shape of South-American Canids: Abiotic or Biotic Processes? Evol Biol 43: 145.Cáceres N., Meloro C., Carotenuto F., Passaro F., Sponchiado J., Melo G.L., Raia P., 2014.

Ecogeographical variation in skull shape of capuchinmonkeys. Journal of Biogeography41: 501–512.

Calder W.A., 1974. Consequences of body size for avian energetics. In: Paynter R.A., Jr(Ed.) Avian energetics. Nuttall Ornithological Club, Cambridge, MA. 86–151.

Canale G.R., Guidorizzi C.E., Kierulff M.C.M., Gatto C., 2009. First record of tool use bywild populations of the yellow-breasted capuchin monkey (Cebus xanthosternos) andnew records for thebearded capuchin (Cebus libidinosus). Am J Primatol. 71(5): 366–372.

Carotenuto F., Diniz-Filho J.A.F., Raia P., 2015. Space and time: The two dimensions ofArtiodactyla body mass evolution. Palaeogeography, Palaeoclimatology, Palaeoecology437: 18–25.

Chaves O.M., Bicca-Marques J.C., 2013. Dietary flexibility of the brown howler monkeythroughout its geographic distribution. American journal of primatology 75(1): 16–29.

Clauss M., Dittmann M.T., Müller D.W.H., Meloro C., Codron D., 2013. Bergmann’s rulein mammals: a cross-species interspecific pattern. Oikos 122: 1465–1472.

Cortés-Ortiz L., Bermingham E., Rico C., Rodrıguez-Luna E., Sampaio I., Ruiz-GarcıaM., 2003. Molecular systematics and biogeography of the Neotropical monkey genus,Alouatta. Molecular phylogenetics and evolution 26(1): 64–81.

Dayan T., Simberloff D., 2005. Ecological and community-wide character displacement:the next generation. Ecology Letters 8: 875–894.

De Lisle S.P., Rowe L., 2015. Ecological character displacement between the sexes. TheAmerican Naturalist 186(6): 693–707.

Diniz-Filho J.A.F., Bini L.M., Hawkins B.A., 2003. Spatial autocorrelation and red herringsin geographical ecology. Global Ecology and Biogeography 12(1): 53–64.

Diniz-Filho J.A.F., Siqueira T., Padial A.A., Rangel T.F., Landeiro V.L., Bini L.M., 2012.Spatial autocorrelation analysis allows disentangling the balance between neutral andniche processes in metacommunities. Oikos 121(2): 201–210.

Dray S., Legendre P., Blanchet G., 2007. Forward selection with permutation (Canoco p.46)(Version 0.0–7). Available at http://r-forge.r-project.org/R/?group_id=195

Eweleit L., Reinhold K., 2014. Body size and elevation: do Bergmann’s and Rensch’srule apply in the polytypic bushcricket Poecilimon veluchianus?. Ecological Entomo-logy 39(1): 133–136.

Fairbairn D.J., 1997. Allometry for sexual size dimorphism: pattern and process in the co-evolution of body size in males and females. Annual review of ecology and systematics659–687.

Fairbairn D.J., 2007. Introduction: the enigma of sexual size dimorphism. In: Fairbairn D.,Blanckenhorn W., Szekely T. (Eds.) Sex, Size and gender Roles. Evolutionary Studiesof Sexual Size Dimorphism. Oxford University Press. 1–10.

Fairbairn D.J., 2013. Odd Couples. Extraordinary Differences Between the Sexes in theAnimal Kingdom. Princeton University Press.

Fairbairn D., BlanckenhornW., Szekely T., 2007. Sex, Size and gender Roles. EvolutionaryStudies of Sexual Size Dimorphism. Oxford University Press.

Ford S.M., 1994. Evolution of sexual dimorphism in body weight in platyrrhines. AmericanJournal of Primatology 34(2): 221–244.

Fragaszy D., Visalberghi E., Fedigan L., 2004. The complete capuchin. Cambridge: Cam-bridge University Press.

Glander K.E., Teaford M.E., 1995. Seasonal, sexual, and habitat effects on feeding rates ofAlouatta palliata. Am Primatol 36: 124–125.

Gohli J., Voje K.L., 2016. An interspecific assessment of Bergmann’s rule in 22mammalianfamilies. BMC evolutionary biology 16(1): 222.

Gordon A.D., 2004. Evolution of body size and sexual size dimorphism in the order Prim-ates: Rench’s rule, quantitative genetics, and phylogenetic effects. PhD dissertation, Uni-versity of Texas, Austin.

Gregory-Wodzicki K.M., 2000. Uplift history of the central and northern Andes: a review.Geol. Soc. Amer. Bullet. 112: 1091–1105.

Griffith D.A., 2013. Spatial autocorrelation and spatial filtering: gaining understandingthrough theory and scientific visualization. Springer Science & Business Media.

Griffith D.A., Peres-Neto P.R., 2006. Spatial modeling in ecology: the flexibility of eigen-function spatial analyses. Ecology 87: 2603–2613.

Harcourt A.H., Schreier B.M., 2009. Diversity, Body Mass, and Latitudinal Gradients inPrimates . Internaional Journal of Primatology 30: 283.

Hijmans R.J., Cameron S.E., Parra J.L., Jones P.G., Jarvis A., 2005. Very high resolutioninterpolated climate surfaces for global land areas. International Journal of Climatology.25: 1965–1978.

Isaac J.L., 2005. Potential causes and life history consequences of sexual size dimorphismin mammals. Mammal Review 35(1): 101–115.

Isaac J.L., Johnson C.N., 2003. Sexual dimorphism and synchrony of breeding: variationin polygyny potential among populations in the common brushtail possum, Trichosurusvulpecula. Behavioral Ecology 14(6): 818–822.

Ives A.R., Midford P.E., Garland T. Jr., 2007. Within-species measurement error in phylo-genetic comparative methods. Systematic Biology 56: 252–270.

126

Rensch and Bergmann’s rules in howler monkeys

James F.C., 1970. Geographic size variation in birds and its relationship to climate. Ecology51: 365–390.

Kay R.F., Plavcan J.M., Glander K.E., Wright P.C., 1988. Sexual selection and canine di-morphism in New World monkeys. American Journal of Physical Anthropology 77(3):385–397.

KelaitaM., Dias P.A.D., Aguilar CucurachiM., Canales Espinosa D., Cortés Ortiz L., 2011.Impact of intrasexual selection on sexual dimorphism and testes size in the Mexicanhowler monkeys Alouatta palliata and A. pigra. American journal of physical anthropo-logy 146(2): 179–187.

Lande R., 1980. Sexual dimorphism, sexual selection, and adaptation in polygenic charac-ters. Evolution 34(2): 292–305.

Lindstedt S.L., Boyce M.S., 1985. Seasonality, fasting endurance, and body size in mam-mals. The American Naturalist 125: 873–878.

Lynch-Alfaro J.W., 2012. How Different Are Robust and Gracile Capuchin Monkeys? AnArgument for the Use of Sapajus and Cebus. Am J Primatol 74(4): 273–286.

Maddison W.P., Maddison D.R., 2011. Mesquite: a modular system for evolutionary ana-lysis, version 2.75. Available at http://mesquiteproject.org

Maestri R., Luza A.L., de Barros L.D., Hartz S.M., Ferrari A., de Freitas T.R.O., DuarteL.D.S., 2016. Geographical variation of body size in sigmodontine rodents depends onboth environment and phylogenetic composition of communities. J. Biogeogr. 43: 1192–1202.

Martinez P.A., Marti D.A., Molina W.F., Bidau C.J., 2013. Bergmann’s rule across theequator: A case study in Cerdocyon thous (Canidae). Journal of Animal Ecology 82:997–1008.

Medina A.I., Martí D.A., Bidau C.J., 2007. Subterranean rodents of the genus Ctenomys(Caviomorpha, Ctenomyidae) follow the converse to Bergmann’s rule. Journal ofBiogeography 34: 1439–1454.

Meiri S., 2011. Bergmann’s Rule — what’s in a name? Global Ecology and Biogeography20(1): 203–207.

Meiri S., Kadison A.E., Novosolov M., Pafilis P., Foufopoulos J., Itescu Y., Raia P.,Pincheira-Donoso D., 2014. The number of competitor species is unlinked to sexualdimorphism. Journal of Animal Ecology 83: 1302–1312.

Meiri S., Yom-Tov Y., Geffen E., 2007. What determines conformity to Bergmann’s rule?.Global Ecology and Biogeography 16: 788–794.

Meiri S., Kadison A.E., Novosolov M., Pafilis P., Foufopoulos J., Itescu Y., PincheiraDonoso D., 2014. The number of competitor species is unlinked to sexual dimorph-ism. Journal of Animal Ecology, 83(6): 1302–1312.

Meloro C., Raia P., Piras P., Barbera C., O’Higgins P., 2008. The shape of the mandibu-lar corpus in large fissiped carnivores: allometry, function and phylogeny. ZoologicalJournal of the Linnean Society 154: 832–845.

Meloro C., Cáceres N., Carotenuto F., Passaro F., Sponchiado J., Melo G.L., Raia P., 2014a.Ecogeographical variation in skull morphometry of howler monkeys (Primates: Atel-idae). Zoologischer Anzeiger – A Journal of Comparative Zoology 253: 345–359.

Meloro C., Cáceres N., Carotenuto F., Sponchiado J., Melo G.L., Passaro F., Raia P., 2014b.In and out the Amazonia: evolutionary ecomorphology in howler and capuchin mon-keys. Evolutionary Biology 41: 38–51.

Pavelka M.S., Knopff K.H., 2004. Diet and activity in black howler monkeys (Alouattapigra) in southern Belize: does degree of frugivory influence activity level? Primates45(2): 105–111.

Peres C.A., 1997. Effects of habitat quality and hunting pressure on arboreal folivore dens-ities in neotropical forests: a case study of howler monkeys (Alouatta spp.). Folia Prim-atologica 68(3–5): 199–222.

Pincheira-Donoso D., 2010. The balance between predictions and evidence and the searchfor universal macroecological patterns: taking Bergmann’s rule back to its endothermicorigin. Theory in Biosciences 129(4): 247–253.

Pinheiro J., Bates D., DebRoy S., Sarkar D., R Core Team, 2016. nlme: Linear and Non-linear Mixed Effects Models. R package version 3.1–128. Available at http://CRAN.R-project.org/package=nlme

Plavcan J.M., 2012. Sexual size dimorphism, canine dimorphism, and male-male compet-ition in primates. Human Nature 23(1): 45–67.

Plavcan J.M., Van Schaik C.P., 1997. Intrasexual competition and body weight dimorphismin anthropoid primates. American Journal of Physical Anthropology 103(1): 37–68.

Rangel T.F., Diniz-Filho J.A.F., Bini L.M., 2010. SAM: a comprehensive application forSpatial Analysis in Macroecology. Ecography 33(1): 46–50.

Ravosa M.J., Ross C.F., 1994. Craniodental Allometry and Heterochrony in Two HowlerMonkeys: Alouatta seniculus and A. palliata. American Journal of Primatology 33:277–299.

Rensch B., 1950. Die Abhängigkeit der relativen Sexualdifferenz von der Körpergrösse.Bonner Zoologische Beiträge. 1: 58–69. [in German]

Revell L.J., 2012. phytools: An R package for phylogenetic comparative biology (andother things). Methods in Ecology and Evolution 3: 217–223.

Rohlf F.J., 2015. The tps series of software. Hystrix 26(1): 9–12. doi:10.4404/hystrix-26.1-11264

Rohlf F.J., Slice D.E., 1990. Extensions of the Procrustes method for the optimal superim-position of landmarks. Systematic Zoology 39: 40–59.

RodríguezM.Á., Olalla-Tárraga1M.A., Bradford A.H., 2008. Bergmann’s rule and the geo-graphy of mammal body size in the Western Hemisphere. Global Ecology and Biogeo-graphy 17: 274–283.

Rosenberger A.L., 1992. Evolution of feeding niches in new world monkeys. Am. J. Phys.Anthropol. 88: 525–562.

Shine R., 1989. Constraints, Allometry, and Adaptation: Food Habits and ReproductiveBiology of Australian Brownsnakes (Pseudonaja: Elapidae). Herpetologica 45(2): 195–207.

Shuster S.M., Wade M.J., 2003. Mating systems and strategies. Princeton University Press.Smith R.J., Cheverud J.M., 2002. Scaling of sexual dimorphism in body mass: a phylogen-

etic analysis of Rensch’s rule in primates. International Journal of Primatology 23(5):1095–1135.

Strier K.B., 1992. Atelinae adaptations: behavioral strategies and ecological constraints.Am. J. Phys. Anthrop. 88: 515–524.

Wallace A.R., 1852. On the monkeys of the Amazon. Proceedings of the Zoological Societyof London. 20: 107–110.

Webb C.O., Ackerly D.D., Kembel S.W., 2008. Phylocom: software for the analysis ofcommunity phylogenetic structure and trait evolution, version 4.0.1. Available at http://www.phylodiversity.net/phylocom

Weckerly F.W., 1998. Sexual-size dimorphism: influence of mass and mating systems inthe most dimorphic mammals. Journal of Mammalogy 79: 33–52.

Werner Y.L., Korolker N., Sion G., Göçmen B., 2016. Bergmann’s and Rensch’s rules andthe spur thighed tortoise (Testudo graeca). Biological Journal of the Linnean Society117(4): 796–811.

Zelditch M.L., Swiderski D.L., Sheets H.D., 2012. Geometric morphometrics for biolo-gists: a primer. Academic Press.

Associate Editor: L.A. Wauters

Supplemental informationAdditional Supplemental Information may be found in the online version of this arti-cle:Figure S1 The 250×250 cell resolution geographical grid used to average morpho-

logical and environmental variables related to the recorded specimens. Redpoints indicate sampling localities, blue points indicate centres of the relatedcells.

Table S2 The dataset used in this study.Figure S3 Phylogenetic tree used in all the phylogenetic informed analyses. The

colour of the branches represents the mapped Sexual Size Dimorhism (SSD).States of internal nodes are reconstructed viaMaximumLikelihood Estimation.

Table S4 Spatial autocorrelation results.Table S5 Results of the regressions between males and females’LnCS and the envir-

onmental variables.

127